Topics in Economic Analysis & Policy

Volume 3, Issue 1 2003 Article 13

National Treatment and the Choice Between

Exports and FDI

Bernard Hoekman∗ Kamal Saggi†

∗
World Bank, bhoekman@worldbank.org
†
SMU, ksaggi@mail.smu.edu

Copyright 2003
c by the authors. All rights reserved. No part of this publication may be
reproduced, stored in a retrieval system, or transmitted, in any form or by any means, elec-
tronic, mechanical, photocopying, recording, or otherwise, without the prior written permission
of the publisher, bepress. Topics in Economic Analysis & Policy is one of The B.E. Jour-
nals in Economic Analysis & Policy, produced by The Berkeley Electronic Press (bepress).
http://www.bepress.com/bejeap.
 National Treatment and the Choice Between
Exports and FDI∗
Bernard Hoekman and Kamal Saggi

Abstract

International trade agreements increasingly constrain the ability of governments to use trade
policies. Fewer international constraints apply to the use of investment policies, although there
is discussion about negotiating such disciplines both regionally and multilaterally. Since firms
compete in foreign markets via both exports and foreign direct investment(FDI), the following
question arises: can constraints on the use of only one type of policy (trade or FDI) induce firms
to adopt inefficient modes of supply when serving foreign markets? We address this question
in a model in which a local and a foreign firm compete in the domestic market. In the model,
the domestic government’s trade and/or FDI policies as well as the foreign firm’s choice between
exports and FDI are endogenous. We show that even if the domestic government is constrained
only in its ability to use trade (FDI) policy, and is free to set its FDI (trade) policy, the foreign firm
chooses the efficient mode of supply. The key point is that it is never in the interest of the domestic
government to set policies in a manner that leads the foreign firm to adopt an inefficient mode of
supply.

KEYWORDS: FDI, Trade Policy, National Treatment, Trade Negotiations, International Agree-
ments

∗
Hoekman: World Bank, 1818 H Street, N.W., Washington, DC 20433. Phone : (202) 473-
1185, fax: (202) 676-9810, e-mail: bhoekman@worldbank.org. Saggi: Department of Eco-
nomics, SMU, Dallas, TX 75275-0496. Phone (214) 768-3274, fax (214) 768-1821, e-mail
ksaggi@mail.smu.edu. We thank two referees, Kyle Bagwell, Philip Levy, Aaditya Mattoo,
Marcelo Olarreaga, and Bob Stern for helpful comments. All errors are our own.
Hoekman and Saggi: Trade vs. Direct Investment 1

1 Introduction
In the last fifty years or so, governments have negotiated numerous agreements that
aim at liberalizing international trade. At the multilateral level, the World Trade Orga-
nization (WTO) greatly limits the discretion of members to raise trade barriers. Both
multilateral and regional trade agreements have also increasingly come to address policies
that indirectly affect trade in order to make market access commitments more secure.
However, much less been done to discipline national policies that affect factor mobility.
Several high income WTO members have argued that there is a need to negotiate mul-
tilateral rules for investment policies, such as the right of establishment and national
treatment for foreign investors. Arguments in support of this position largely revolve
around market access objectives: in many sectors the preferred mode of contesting a
market may be through foreign direct investment (FDI), not exports. Under such a
scenario, if rules exist on trade policy but not on investment policy, one may conjecture
that measures imposed by governments in the latter area may distort the choice of mode
of supply of foreign firms. In fact, such concerns have led to policy proposals in favor
of modal neutrality.1 That is, the suggestion is that rules should be designed to ensure
that government policies do not lead to the choice of an inefficient mode of supply.2
In this note we analyze the relationship between trade and FDI policy in a market in
which a domestic and a foreign firm compete in the domestic market.3 Our main goal is
to assess the potential efficiency impact of international agreements that constrain the
ability of governments to use trade policy and/or taxes that discriminate against foreign
investors (i.e. taxes that violate the principle of national treatment). The former scenario
is of interest because it represents the status quo: WTO members face constraints on the
use of trade policy (due to commitments on tariff bindings and general WTO disciplines),
but not (yet) on investment policy. The latter issue is relevant since a national treatment
requirement for foreign investors is likely to be a central element of any multilateral
investment agreement. We compare a situation where the domestic government is free
to use two policy instruments (one each for trade and FDI) with a situation where
the ability to use one of the two instruments is constrained. There are two interesting
questions here: (1) Can constraints on the use of only one type of policy (trade or FDI)
induce the foreign firm to adopt an inefficient mode of supply? (2) What is the marginal
value of constraining FDI policy given that trade policy is already constrained?4
1
See for example Hoekman (1996), Feketekuty (2000), and Low and Mattoo (2000).
2
Another line of argument emphasizes the potential payoffs to developing countries of signing on to
multilateral rules as a commitment device. See Markusen (2001) and Moran (1998) for discussions of
the potential interests of developing countries in negotiating multilateral rules on investment policies.
3
Since multinational firms are found mostly in oligopolistic industries, we develop an oligopoly model
in which firms compete in quantities. Our approach is similar to Levinsohn (1989) and Horstmann and
Markusen (1992).
4
See Hoekman and Saggi (2000) for a general discussion of the case for a multilateral investment

Produced by The Berkeley Electronic Press, 2004

2 Topics in Economic Analysis & Policy Vol. 3 [2003], No. 1, Article 13

As is well known, in models such as ours, tariffs on foreign firms create a globally
inefficiency if costs of production of local and foreign firms are similar in magnitude.5
What has not received adequate attention, however, is whether an asymmetric policy
constraint may force the foreign firm to choose an inefficient mode of supply. For ex-
ample, at an intuitive level, it seems conceivable that a tax on a foreign firm’s local
production may compel it to service the market through exports, when actually FDI is
more efficient (in the absence of such a tax). Alternatively, it seems plausible that a
constraint on the use of FDI policy in the presence of zero tariffs might bias the choice
of firms towards FDI viz-a-viz trade.
Our simple model shows that the logic behind above concerns is incomplete. In fact,
in the equilibrium of our model, such distortions do not arise even when a discriminatory
output (trade) tax is imposed on the foreign firm and trade (FDI) is subject to no such
tax. In the case where investment policy is not constrained (the scenario that is more
relevant for the actual policy debate), the interesting finding is that despite the lack of
national treatment for FDI, the foreign firm’s choice between different modes of supply
does not get distorted: it chooses the efficient mode. The key mechanism driving this
result is the endogenity of policy in our model: if one were to compare a scenario with
arbitrary taxes on FDI and zero tariffs, inefficiency with respect to mode choice can
obviously be obtained. However, our model argues that this is an incorrect comparison
since taxes that induce such inefficiencies may actually never arise in equilibrium.
Although the absence of national treatment does lead to an inefficient level of out-
put relative to laissez-faire if firms are relatively symmetric in terms of their costs of
production, this distortion is mild in the sense that the equilibrium output under FDI
subject to a tax is no lower than that which would obtain were the foreign firm to export
under a zero tariff regime. Thus, aggregate world welfare under the optimal FDI tax is
identical to that which obtains under free trade. Our analysis suggests that the primary
benefit of constraining FDI policy given that trade is free is primarily distributional:
such a constraint would limit the ability of governments to extract rents from foreign
investors.

2 Model
We restrict attention to the domestic market that produces two goods: x and y. Pref-
erences over these goods are quasi-linear: U(x, y) = u(x) + y where y is the numeraire
good produced under perfect competition with constant returns to scale technology. As
is well known, the above preference structure implies that the demand for good x is a
function of only its price (i.e. our model is a partial equilibrium one). Let p(x) denote
agreement at the WTO.
5
See Brander and Spencer (1984) and the large literature on strategic trade policy (surveyed in
Brander, 1995).

http://www.bepress.com/bejeap/topics/vol3/iss1/art13
Hoekman and Saggi: Trade vs. Direct Investment 3

the inverse demand function for good x.

Good x is produced by a domestic firm that competes with a foreign firm. We study
a three stage game between the two firms and the domestic government. In the first
stage of the game, the government chooses its policy stance: it announces the specific
import tariff t and/or an output tax τ that is imposed on the foreign firm should it
decide to produce in the domestic market via foreign direct investment (FDI). In the
second stage, the foreign firm chooses its mode of supply (exports or FDI). In the final
stage, the two firms compete as Cournot-Nash competitors.
We assume that FDI policy is an output tax for ease of exposition: the comparison
between an import tariff and an output tax is more transparent than the comparison
between an import tariff and alternative FDI policies such as discriminatory profit taxes,
equity ownership limits on foreign investors, locational restrictions, and cost-raising
policies such as local content requirements for either intermediates inputs or factors
of production.6 Furthermore, since governments can intervene in both trade and FDI
via multiple instruments, it is reasonable to focus on two instruments that have similar
effects on the market equilibrium.
The domestic firm’s marginal cost of production equals cH whereas that of the foreign
firm equals cF (under FDI) and φcF (under exports). In general, exporting carries the
disadvantage of transportation costs whereas FDI requires the foreign firm to operate
in a relatively unfamiliar environment. As a result, either mode of supply may be more
efficient from the foreign firm’s perspective and the parameter φ captures the efficiency
difference between the two modes. When φ = 1 the two modes are equally efficient and
if φ < (>) 1, exports (FDI) is the more efficient (less costly) mode. Of course, exporting
and FDI may differ in several other ways: for example, FDI may generate technological
spillovers for the local firm (see Saggi, 2002). Our model abstracts from such issues and
focuses only on static efficiency differences between the two modes of supply. 7
The output tax facing the domestic firm is normalized to zero in order to facilitate
the comparison between trade policy and FDI policy. Thus the FDI tax is by definition
discriminatory. If the government chooses a zero import tariff, it is said to commit to
free trade in the first stage of the game. Similarly, if it commits to a zero output tax
on the foreign firm under FDI, it is said to commit to national treatment of the foreign
firm (the tax facing the domestic firm is zero).
6
While some of these policies are equivalent to an output tax, others (such as ownership restrictions)
are not. However, even policies that do not behave like an output tax will affect our results in a
straightforward fashion. A detailed investigation of such policies is left for future research.
7
If governments believe there are positive spillovers associated with FDI and therefore skew policy
towards attracting FDI relative to exports, the results below will require modifications.

Produced by The Berkeley Electronic Press, 2004

4 Topics in Economic Analysis & Policy Vol. 3 [2003], No. 1, Article 13

3 Analysis
To solve for a sub-game perfect Nash equilibrium, we begin at the last stage of the
game. At this stage, the two firms compete in quantities, given the policies set by the
government and the foreign firm’s choice between exports and FDI. Let i = H, F denote
the identity of a firm and j = E, F denote the foreign firm’s mode of supply (E denotes
exports and F denotes FDI). The marginal cost of firm i under mode j is denoted by cji
where the foreign firm’s cost is inclusive of any tariff (t) or tax (τ ) that it faces.8
Let xj ≡ xjH + xjF be the total output sold in the domestic market and pj the
associated equilibrium price under mode j. Define aggregate welfare in the host country
under mode j by:
W j ≡ πjH + CS j + rxjF (1)
where πjH = (pj − cH )xjH denotes the domestic firm’s profit; CS j = u(xj ) − pj xj the total
consumer surplus in the domestic market; r = τ or t the relevant government policy;
and rxjF the total tax/tariff revenue.
Given the cost structure assumed, it is obvious that absent any policy intervention,
the foreign firm opts for exports iff φ ≤ 1. To understand the interaction between the
two types of policies and to analyze equilibrium policies, consider each policy in isolation.

3.1 Trade Policy

Define t∗ as the domestic economy’s optimal import tariff, given that exporting is the
only mode of supply available to the foreign firm. We have

t∗ ≡ Arg max W E (t) ≡ π E E E

H (t) + CS (t) + txF (t) (2)

As Brander and Spencer (1984) have shown, such a tariff is in general positive and it
balances the incentives for rent extraction from the foreign firm against the interests of
consumers.
Since the foreign firm can always produce in the domestic market by opting for FDI,
the domestic government may face a constraint on its choice set. Define tF as the tariff
that just deters FDI on the foreign firm’s part (as in Konishi et. al., 1999). Since the
foreign firm’s marginal cost under exports equals φcF + t whereas under FDI it equals
cF , we have:
tF = (1 − φ)cF (3)
Note that if t∗ > tF the domestic government is incapable of implementing its optimal
tariff (any tariff higher than tF leads to tariff jumping FDI). We assume this condition
8
We ignore the possibility of subsidies here. Note, however, that the tariff or the output tax on
the foreign firm results in rent shifting much like a subsidy to the domestic firm, so that we need not
consider subsidies in addition to these two policies.

http://www.bepress.com/bejeap/topics/vol3/iss1/art13
Hoekman and Saggi: Trade vs. Direct Investment 5

holds or else the option of FDI does not have any interesting implications for trade
policy. The following lemma notes an important property of the tariff tF :
Lemma 1: Domestic welfare gross of tariff revenue (i.e. the sum of consumer surplus
and domestic profits) is identical under FDI and exports subject to the tariff tF .
The logic behind this lemma is simple. Market equilibrium is identical under exports
and FDI so long as the costs of both firms are the same under the two modes of sup-
ply. The tariff tF implies that the foreign firm’s marginal cost under both exports and
FDI equals cF . The domestic firm’s cost is cH , regardless of the foreign firm’s mode of
supply. Thus, equilibrium outputs of the two firms (and therefore price, profits, and con-
sumer surplus) are identical under exports and FDI. We next describe the government’s
equilibrium trade policy:
Proposition 1: Suppose (i) t∗ > tF ; (ii) φ ≤ 1; and (iii) the tariff is the only policy
instrument available. Then, the government implements the tariff tF that just deters
FDI by the foreign firm (i.e. domestic welfare is higher under exports subject to the tariff
tF than under FDI with national treatment).9
Why does not the government impose a tariff higher than tF thereby inducing FDI
by the foreign firm? The reason is that when φ ≤ 1, exporting is a more efficient mode
of supply than FDI and the domestic government can collect the cost savings the foreign
firm enjoys relative to FDI via the tariff tF while at the same time ensuring that domestic
consumers and the local firm are no worse off than they are under FDI (see lemma 1).
A similar intuition applies in the case where the government is free to choose only its
FDI policy.
3.2 FDI Policy
As in the case of trade policy, define τ ∗ as the optimal output tax on the foreign firm
under FDI, given that FDI is the only method of serving the market. We have10

τ ∗ ≡ Arg max WHF (τ ) ≡ π FH (τ ) + CS F (τ ) + τ xFF (τ ) (4)

The first order condition for the above problem is completely analogous to the case of
the tariff since an import tariff is simply a tax on the output sold by the foreign firm
via exports. As in the case of the tariff tF , define τ F as the output tax that just deters
exports on the foreign firm’s part (any tax less than or equal to τ F makes FDI more
profitable than exports). Recall that the foreign firm’s marginal cost under exports
equals φcF whereas under FDI it equals τ + cF . Therefore,

τ F = (φ − 1)cF (5)
9
Note that if φ > 1, the domestic government is incapable of implementing any positive tariff.
10
We assume that all profits of the foreign firm, net of the tax under FDI, are repatriated and do not
contribute to domestic welfare.

Produced by The Berkeley Electronic Press, 2004

6 Topics in Economic Analysis & Policy Vol. 3 [2003], No. 1, Article 13

We can then show a result very similar to proposition 1. Suppose (i) τ ∗ > τ F ; (ii) φ ≥ 1;
(iii) and the output tax (τ ) on local production by the foreign firm is the only policy
instrument available. Then, the government implements the tax τ F that just deters the
foreign firm from switching to exports (i.e. domestic welfare is higher under FDI subject
to the tax τ F than under free trade). The intuition is the same as before: it is optimal
for the government to ensure that the foreign firm adopts the more efficient mode of
supply so that it can tax the foreign firm’s cost savings relative to the inefficient mode.
Note that we do not need τ ∗ > τ F . When this inequality is reversed the foreign firm still
adopts the efficient mode of supply (it does FDI only if τ ∗ + cF < φcF ) and constraining
the local government’s investment policy again has mainly distributional consequences.
Thus, in a situation where only the use of tariffs is constrained (by a trade agreement
say), taxation of local production by the foreign firm does not distort its choice between
alternative modes of serving the local market. Our main insight is that the tax is never
set so high as to distort the modal choice, despite the profit-shifting interest of the
government. The reason is that the domestic government has an incentive to induce the
mode under which the tax rate is higher. This is because the more efficient the foreign
firm’s chosen mode of supply, the higher the tax the local government can implement
holding constant the welfare of all other local agents.
3.3 Trade and FDI Policy
Consider now the equilibrium of the full game where the government is free to choose
both its trade and FDI policy (t and τ ). As before, the last stage simply involves Cournot
competition, so we skip its discussion. Consider the foreign firm’s decision regarding its
mode of supply. Given the output tax τ and the tariff t, the foreign firm opts for exports
over FDI iff
t + φcF ≤ τ + cF ⇔ t ≤ t(τ h ) ≡ τ + (1 − φ)cF (6)
It is clear from (6) that the two policies (t and τ ) are complementary: t(τh ) is increasing
in τ . In other words, a higher output tax on the foreign firm under FDI allows the
government to implement a higher tariff. Note also that if t = τ , the foreign firm opts for
exports iff φ ≤ 1: in this scenario, government policy imposes the same cost disadvantage
under both exports and FDI, thereby leaving the decision to be determined solely by
the true relative cost comparison. The equilibrium policy pair (t∗ , τ ∗ ) is described in
the following proposition:
Proposition 2: When the domestic government has both policy tools at its disposal,
its equilibrium policy is as follows:
(i) If φ = 1, equilibrium policy is the pair {t∗ , τ ∗ } and t∗ = τ ∗ .
(ii) If φ < 1, any pair {t∗ , τ }, where τ > τ and τ ≡ t∗ −(1−φ)cF , is an equilibrium
policy and t∗ > τ .
(iii) If φ > 1, any pair {t, τ ∗ } where t > t and t ≡ τ ∗ − (φ − 1)cF is an equilibrium

http://www.bepress.com/bejeap/topics/vol3/iss1/art13
Hoekman and Saggi: Trade vs. Direct Investment 7

policy and τ ∗ > t .

When FDI and exports are equally efficient (i.e. φ = 1), the domestic government is
able to implement local welfare maximizing tariff and tax levels (t∗ and τ ∗ ). When faced
with the policy pair {t∗ , τ ∗ }, where t∗ = τ ∗ , the foreign firm is indifferent between the
two modes of supply. Since both modes are equally efficient from the social viewpoint,
policy intervention does not result in the foreign firm adopting an inefficient mode of
supply.
When φ < 1, exporting is more efficient than FDI in the absence of policy interven-
tion. For a policy pair {t∗ , τ }, where τ > τ , to not result in an inefficient mode of supply,
the foreign firm should choose to export when facing any such pair. Indeed it does. The
reason is clear: since τ > t∗ − (1 − φ)cF , it must be that φcF + t∗ < cF + τ . When
exporting is more efficient, domestic welfare is higher under an optimal import tariff
than under FDI subject to the optimal output tax. The reason is that a more efficient
mode of supply allows the domestic government to extract higher revenue. Consequently,
the domestic government implements an output tax that is high enough to support its
optimal tariff while at the same time ensuring that the foreign firm indeed chooses to
export. The interpretation of part (iii) of the above proposition is analogous to part
(ii).
As should be expected, having two instruments at its disposal does alter the outcome
in favor of the domestic government since it can implement a tariff (tax) as high as it can
under the scenario where the foreign firm does not have the option of exporting (FDI).
What the analysis above informs us is that, regardless of whether the government has
only one or both policy instruments at its disposal, policy intervention does not result
in the foreign firm adopting an inefficient mode of supply even though the two modes
may differ significantly in their welfare effects.11

4 Concluding remarks
Currently, most countries are free to chose their FDI policies whereas tariffs are bound by
the WTO. In the case of free trade agreements or customs unions, trade between member
countries can be subject to no tariffs at all. Our analysis implies that such selective
policy constraints may not necessarily distort the decisions of firms from non-member
countries regarding their preferred means of serving member country markets. Moreover,
given that trade policies are constrained by a trade agreement, adding constraints on
FDI policies of member countries may primarily have distributional consequences: such
11
As in Brander and Spencer (1984), it is easy to show that world welfare (sum of foreign firm’s
profits and aggregate domestic welfare) declines due to domestic policy intervention when the firms
have relatively equal marginal costs of production. When the foreign firm is relatively inefficient, a
tariff or a tax under FDI can improve world welfare by allocating a greater share of the world output
to the lower cost (domestic) firm.

Produced by The Berkeley Electronic Press, 2004

8 Topics in Economic Analysis & Policy Vol. 3 [2003], No. 1, Article 13

countries will lose their ability to extract rents from foreign firms without affecting
firm choices regarding mode of supply. This is not to say that policy intervention with
respect to FDI can never have efficiency consequences. In fact, even in our model, such
intervention creates an inefficiency in that too little output is produced. When firms
are equally efficient, an output tax on the foreign firm lowers world welfare. Our main
point is that it is in the interest of a host country to ensure that foreign firms adopt
the truly efficient mode of supply. Thus, policy intervention should not be expected to
distort firm decisions regarding modes of supply.
Our simple model abstracts from several important considerations. For example,
governments might incur higher costs in collecting discriminatory FDI taxes relative to
collecting tariffs on imports. Under such a scenario, governments may have an incentive
to set a high FDI tax so as to induce the foreign firm to export even if FDI is the more
efficient mode of supply. In general, there are a variety of real world circumstances that
may induce governments to set policy so as to induce the choice of an inefficient mode
of supply by the foreign firm. However, our objective here is to simply to show that,
abstracting from such considerations, a constraint on the use of tariffs or FDI policies
does not necessarily distort the choice of mode of supply. Whether our argument holds
in a multi-country world or under alternative assumptions regarding market structure
and firm behavior is worthy of future research.

References
[1] Brander, J. and B. Spencer 1984. “Tariff protection and imperfect competition,”
in ed. H. Kierzkowski Monopolistic Competition and International Trade, Oxford
University Press.
[2] Brander, J. 1995. “Strategic trade policy,” in the Handbook of International Eco-
nomics, vol. 3, edited by G. M. Grossman and K. Rogoff, Elsevier Science Publish-
ers, 1995.
[3] Feketekuty, G. 2000. “Assessing and improving the architecture of GATS,” Chapter
4 in Sauve, P. and Stern, R. (eds.), Services 2000: New Directions in Services Trade
Liberalization, Brookings Institution and Harvard University, Washington D.C.
[4] Hoekman, B. 1996. “Assessing the general agreement on trade in services.” Chapter
4 in Martin, W. and Winters, A (eds.), The Uruguay Round and the Developing
Countries, Cambridge University Press, Cambridge.
[5] Hoekman, B. and K. Saggi 2000. “Assessing the Case for Extending WTO Disci-
plines on Investment-Related Policies,” Journal of Economic Integration 15: 629-
653.
[6] Horstmann, I. J. and J.R. Markusen 1992. “Endogenous market structures in inter-

http://www.bepress.com/bejeap/topics/vol3/iss1/art13
Hoekman and Saggi: Trade vs. Direct Investment 9

national trade,” Journal of International Economics 32: 109-129.

[7] Konishi, H., K. Saggi, and S. Weber 1999. “Endogenous trade policy under FDI,”
Journal of International Economics 49: 289-308.
[8] Levinsohn, J. A. 1989. “Strategic trade policy when firms can invest abroad: when
are tariffs and quotas equivalent?” Journal of International Economics 27: 129-146.
[9] Low, P. and Mattoo, A. 2000. “Is there a better way? Alternative approaches to
liberalization under the GATS,” Chapter 15 in Sauve, P. and Stern, R. (eds.), Ser-
vices 2000: New Directions in Services Trade Liberalization, Brookings Institution
and Harvard University, Washington D.C.
[10] Markusen, J. 2001. “Multilateral rules on foreign direct investment: The developing
countries’ stake,” Review of International Economics 9(2): 287-302.
[11] Moran, T. 1998. Foreign direct investment and development, Washington DC, In-
stitute for International Economics.
[12] Saggi, K. 2002. “Trade, foreign direct investment, and international technology
transfer: a survey,” World Bank Research Observer 17: 191-235.

Produced by The Berkeley Electronic Press, 2004

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.